There are 2 versions of this ad. Below is the anodyne version. Bonus points if you can find the more analeptic version. And provide proof of your quest. Good luck!
* You will work for a fast-growing AI company in London, building language models that help real people in real world.
* You will preprocess and clean large text datasets, removing noise and polishing tokens until the patterns shine.
* You will experiment with embeddings, attention layers, and prompt designs to unlock hidden insights from language.
* You will fine-tune pre-trained LLMs on domain-specific data, making them more accurate, responsive, and context-aware.
* You will evaluate models using metrics like perplexity, BLEU, ROUGE, F1 score, and user-centric feedback.
* You will implement prompt engineering strategies to ensure outputs are high-quality, reliable, and safe.
* You will help integrate LLMs and other ML models into applications, APIs, and production pipelines.
* You will monitor performance and debug models to identify hallucinations, biases, and unexpected outputs.
* You will collaborate with data engineers, NLP researchers, software developers, and product owners to ensure models meet real-world needs.
* You will document models, datasets, and experiments, so knowledge can be reused and refined.
* You will use version control systems (e.g. Git) to keep your code polished and traceable.
* You will be comfortable with CI/CD pipelines for model deployment so your LLMs are production-ready.
* You will learn and apply ethical AI and ML best practices, ensuring models are responsible, fair, and explainable.
* You will explore new libraries, frameworks, and tools to sharpen your skills: transforming rough data into dazzling insights.
Still with us? Great. Then please send us a CV and a covering letter convincing us that you actually enjoyed reading the above. We look forward to hearing from you.